摘要
引入任意随机变量序列极限对数似然比概念,作为任意相依随机序列联合分布与其边缘乘积分布"不相似"性的一种度量,利用截尾方法构造几乎处处收敛的上鞅,给出了一个关于任意随机序列的强偏差定理。
The notion of limit relative logarithmic likelihood ratio of stochastic sequences, as a measure of 'dissimilarity' between their joint distributions and the product of their marginals, is introduced. Construct convergence super-martingale by means of truncation method.A strong deviation theorem for arbitrary stochastic sequence is obtained.
出处
《安徽工业大学学报（自然科学版）》
CAS
2005年第3期308-309,共2页
Journal of Anhui University of Technology（Natural Science）
基金
安徽省教育厅自然科学研究基金项目(2004Kj069)
关键词
随机序列
极限相对对数似然比
强偏差定理
random sequence
limit relative logarithmic likelihood ratio
strong deviation theorem
